files to reestimate the data examples in the following article:

Walter R. Mebane, Jr. and  Jasjeet Singh Sekhon. 2004.  ``Robust Estimation
and Outlier Detection for Overdispersed Multinomial Models of Count Data.'' 
American Journal of Political Science 48 (April): 391--410.
http://macht.arts.cornell.edu/wrm1/multinom.pdf

This file was prepared by
Walter R. Mebane, Jr. and Jasjeet Singh Sekhon
5 March 2004

The files are as follows.

  -rw-rw----    1 wrm1     research     1305 Mar  5 00:29 FLfull1.a.R
  -rw-rw----    1 wrm1     research   109914 Mar  5 01:05 FLfull1.a.RData
  -rw-rw----    1 wrm1     research    41302 Mar  5 01:05 FLfull1.a.Rout
  -rw-rw----    1 wrm1     research     1365 Mar  5 01:38 FLfull1.b.R
  -rw-rw----    1 wrm1     research   109914 Mar  5 03:14 FLfull1.b.RData
  -rw-rw----    1 wrm1     research    49844 Mar  5 03:14 FLfull1.b.Rout
  -rw-rw----    1 wrm1     research     1196 Feb 21 23:26 FLfull1.R
  -rw-rw----    1 wrm1     research   220276 Feb 21 23:52 FLfull1.RData
  -rw-rw----    1 wrm1     research    48649 Feb 21 23:52 FLfull1.Rout
  -rw-rw----    1 wrm1     research      377 Feb 22 00:12 FLsum.R
  -rw-rw----    1 wrm1     research    13731 Feb 22 00:14 FLsum.Rout
  -rw-rw----    1 wrm1     research     2264 Mar  4 15:43 jj2.R
  -rw-rw----    1 wrm1     research   437567 Mar  4 16:46 jj2.RData
  -rw-rw----    1 wrm1     research   130126 Mar  4 16:46 jj2.Rout
  -rw-rw----    1 wrm1     research     2127 Mar  4 21:32 jj.a.R
  -rw-rw----    1 wrm1     research   198010 Mar  4 23:17 jj.a.RData
  -rw-rw----    1 wrm1     research    78563 Mar  4 23:17 jj.a.Rout
  -rw-rw----    1 wrm1     research     2188 Mar  5 15:07 jj.bb.R
  -rw-rw----    1 wrm1     research   198010 Mar  5 17:05 jj.bb.RData
  -rw-rw----    1 wrm1     research    53422 Mar  5 17:05 jj.bb.Rout
  -rw-rw----    1 wrm1     research     2183 Mar  5 15:05 jj.b.R
  -rw-rw----    1 wrm1     research   198010 Mar  5 16:29 jj.b.RData
  -rw-rw----    1 wrm1     research    44936 Mar  5 16:29 jj.b.Rout
  -rw-rw----    1 wrm1     research     1976 Mar  4 15:36 jj.R
  -rw-rw----    1 wrm1     research   381314 Mar  4 17:07 jj.RData
  -rw-rw----    1 wrm1     research    53014 Mar  4 17:07 jj.Rout
  -rw-rw----    1 wrm1     research    12103 Jul 25  2002 pareplic.asc
  -rw-rw----    1 wrm1     research     4144 Mar  5 17:17 README
  -rw-rw----    1 wrm1     research    32264 Jul 23  2002 REPLICATION_DATA.asc
  -rw-rw----    1 wrm1     research     4020 Jul 23  2002 REPLICATION_DESCRIPTION

FLfull1.R:  estimates the model shown in Table 5 of the paper (Florida).  the
results using the multinomRob() program with the default arguments for rgenoud()
are slightly better (in terms of the size of the LQD estimate of the residual
dispersion) than the ones that appear in the paper.  the estimation results are
saved in FLfull1.RData

FLfull1.a.R, FLfull1.b.R:  runs that use the results from FLfull1.R to try to
find a better solution.

FLsum.R:  prints a summary of the results from FLfull1.R with the county names
included as labels.

jj.R:  estimates the model shown in Table 7 of the paper (Poland).  the results
using the multinomRob() program with the default arguments for rgenoud() are
slightly worse (in terms of the size of the LQD estimate of the residual
dispersion) than the ones that appear in the paper.  the estimation results are
saved in jj.RData

jj2.R:  uses equality constraints to reformulate the Poland data model to allow
starting values to be obtained from the multinomial-t model.  fixing the
multinomial-t DF parameter is necessary to get estimates from that model, which
in any case appear to be inferior to the simple multinomial model estimates.

jj.a.R, jj.b.R, jj.bb.R:  runs that use the results from jj.R to try to find a
better solution.

REPLICATION_DATA.asc, REPLICATION_DESCRIPTION:  Florida replication data set

pareplic.asc:  Poland replication data set (kindly provided by John Jackson).

NOTE:  the programs were run on a variety of computers with greatly varying
capabilities, so the run times shown at the end of the various .Rout files are
not necessarily comparable.
